Skip to content

beginnerguide.md

Latest commit

 

History

History
84 lines (52 loc) · 3.58 KB

beginnerguide.md

File metadata and controls

84 lines (52 loc) · 3.58 KB

Beginner's Guide

This guide presents three different methods of accessing S3DF. This is an example of a general step-by-step workflow we hope is suitable for most users. Within this document, you will find guidance on how to:

  • Log in to the S3DF system
  • Navigate directories and storage spaces
  • Access supported applications
  • Prepare and submit a job script

Follow these instructions to access and use S3DF.

Let's get started!

Access to S3DF Through SSH

This example provides a clear, step-by-step workflow for running software on S3DF through SSH.

Connect to a Bastion Node

To start, connect to a bastion node using the following command:

ssh username@s3dflogin-mfa.slac.stanford.edu

Connect to an Interactive Node

After successfully connecting to a bastion node, log in to an interactive node using SSH. For example:

ssh iana

Set Up a Running Environment

To set up your running environment, create a bash file containing all necessary commands, and then execute the bash file.

Configure an SLURM Job Script

Here is an example SLURM job script named run.sbatch:

               #!/bin/bash
              #SBATCH --partition=milano
              #SBATCH --account=rfar
              #SBATCH --job-name=test
              #SBATCH --output=output-%j.txt
              #SBATCH --error=error-%j.txt
              #SBATCH --nodes=1
              #SBATCH --ntasks-per-node=16
              #SBATCH --time=0-00:10:00
              mpirun /sdf/group/rfar/ace3p/bin/omega3p pillbox.omega3p
  • Submit Jobs to a Compute Node

Use the sbatch command to submit your job to a compute node for execution:

              sbatch run.sbatch
  • Check the Status of Running Jobs (Optional)

To monitor the status of your submitted jobs, run the following command:

              squeue -u username
  • View Data Output

Once your jobs have completed, you can view the data output directly on the pool node to verify that the results are as expected.

  • Transfer Data (If Necessary)

If you need to transfer data, connect to a data transfer node to facilitate the movement of your files. Use appropriate file transfer commands (e.g., scp, rsync) to move your data to the desired location.

Access to S3DF Through NoMachine

  • NoMachine offers a specialized remote desktop solution that enhances the performance of X11 graphics over slow connections, compared to SSH.
  • A key feature of NoMachine is its ability to maintain the state of your desktop across multiple sessions, even if your internet connection is unexpectedly lost.
  • To access NoMachine, use the login pool at s3dfnx.slac.stanford.edu.
  • For additional information about this access method, please refer to the NoMachine documentation.

Access to S3DF Through OnDemand

  • Users can also access S3DF through Open OnDemand via any (modern) browser.
  • This solution is recommended for users who want to run Jupyter notebooks, or don't want to learn SLURM, or don't want to download a terminal or the NoMachine remote desktop on their system.
  • After login, you can select which Jupyter image to run and which hardware resources to use (partition name and number of hours/cpu-cores/memory/gpu-cores).
  • The partition can be the name of an interactive pool or the name of a SLURM partition.
  • You can choose an interactive pool as partition if you want a long-running session requiring sporadic resources; otherwise slect a SLURM partition.
  • Note that no GPUs are currently available on the interactive pools.